---
title: "Moral Values Across Countries"
output: 
  flexdashboard::flex_dashboard:
    orientation: columns
    vertical_layout: fill
    source_code: embed
    theme: united
---

```{r setup, include=FALSE}
# global options
knitr::opts_chunk$set(echo = FALSE,
                      tidy = TRUE,
                      cache = FALSE,
                      message = FALSE, 
                      error = FALSE, 
                      warning = FALSE)

# packages
library(flexdashboard)
library(here)
library(rio)
library(tidyverse)
library(magrittr)
library(lme4)
library(lmerTest)
library(dotwhisker)
library(tidytext)
library(ggeffects)
library(htmlwidgets)
library(gt)
library(gghighlight)

theme_set(theme_minimal()) # set theme
options(scipen = 999) # remove scientific notation
```

```{r wrangling, include = FALSE}
# import data
df <- import(here("data", "ALL_MFQ30.csv"),     # moral values, countries, & sex
             setclass = "tbl_df") %>% 
  janitor::clean_names()

df_c <- import(here("data", "Data_S1_sec.csv"), # country-level variables
               setclass = "tbl_df") %>% 
  janitor::clean_names()

# data wrangling
df %<>% 
  drop_na() %>%                             
  mutate(
    across(where(is.double), as.numeric),   
    across(where(is.character), as.factor), 
    
    sex = recode(sex,                       
                 `1` = "Male",
                 `0` = "Female",
                 .default = NA_character_),
    
    indiv = rowMeans(                
      select(., harm_avg, fairness_avg) # individualizing moral foundations
    ),
    
    bind = rowMeans(                 
      select(., ingroup_avg:purity_avg) # binding moral foundations
    )
  )

# check data structure and variables
str(df)

# descriptive statistics by country
c_desc <- 
df %>% 
  pivot_longer(cols = c(indiv, bind),
               names_to = "vars",
               values_to = "val"
               ) %>% 
  select(country, vars, val) %>% 
  group_by(country, vars) %>% 
  summarise(mean = mean(val, na.rm = TRUE),
            sd = sd(val, na.rm = TRUE),
            min = min(val, na.rm = TRUE),
            max = max(val, na.rm = TRUE),
            .groups = "drop"
            ) %>% 
  mutate(vars = fct_recode(vars,
                           Individualizing = "indiv",
                           Binding = "bind"
                           )
         )

# descriptive statistics by country and sex
c_s_desc <- 
df %>% 
  filter(country != "Poland") %>%            # Poland has missing data in sex.
  pivot_longer(cols = c(indiv, bind),
               names_to = "vars",
               values_to = "val"
               ) %>% 
  group_by(country, sex, vars) %>% 
  summarise(mean = mean(val, na.rm = TRUE),
            sd = sd(val, na.rm = TRUE),
            min = min(val, na.rm = TRUE),
            max = max(val, na.rm = TRUE),
            .groups = "drop"
            ) %>% 
  mutate(vars = fct_recode(vars,
                           Individualizing = "indiv",
                           Binding = "bind"
                           )
         )
```


# Values X Country

Background {.sidebar}
-----------------------------------------------------------------------
**Data**  
Data used in this dashboard come from the second study of [Atari et al. (2020)](http://dx.doi.org/10.1098/rspb.2020.1201), which can be downloaded from [Kaggle](https://www.kaggle.com/tunguz/sex-differences-in-moral-judgements-67-countries).  

This study has data on moral values in 19 countries. There is also a second dataset with country-level variables.

**Country-level variables**  
  
Population Sex Ratio  
Individualism  
Masculinity  
Gender Equality  
Human Development Index  
Overall Life Satisfaction Index  

**Definitions**

Moral foundations theory [(Graham et al., 2013)](https://doi.org/10.1016/B978-0-12-407236-7.00002-4) argues that there are five universal moral values. These five values have been grouped into two overarching moral values [(Graham et al., 2009)](http://dx.doi.org/10.1037/a0015141) as individualizing and binding moral values.  

*Individualizing* values encompass moral foundations of care (i.e., cherishing and protecting others) and fairness (i.e., rendering justice according to shared rules).  

*Binding* values encompass moral foundations of loyalty (i.e., standing with your group, family, nation), authority (i.e., submitting to tradition and legitimate authority), and purity (i.e., abhorrence for disgusting things, foods, actions).  

**Example Items**  

Please read the following sentences and indicate your agreement or disagreement:  

Individualizing: Compassion for those who are suffering is the most crucial virtue.  

Binding: People should be loyal to their family members, even when they have done something wrong.  

Column {data-width=600}
-----------------------------------------------------------------------

### **Final Version**

```{r}
gmeans <- 
c_desc %>% 
  group_by(vars) %>% 
  summarise(m = mean(mean))

c_desc %>% 
  ggplot() + 
  geom_vline(
    data = gmeans,
    aes(xintercept = m), 
    linetype = 2,
    alpha = .6
            ) + 
  geom_vline(
    data = data.frame(x = seq(0,5,1)),
    aes(xintercept = x),
        color = "gray80",
        alpha = .6
             ) +
  geom_hline(aes(yintercept = 0.5),
             alpha = .6,
             color = "gray80") +
  geom_col(
    aes(mean, reorder_within(country, mean, vars), 
        fill = country
        ),
    alpha = .7,
    position = position_nudge(y = -0.05)
          ) + 
  scale_y_reordered() + 
  scale_x_continuous(expand = c(0, 0.32),
                     limits = c(0, 5),
                     breaks = c(0, 5),
                     labels = c("Strongly\nDisagree",
                                "Strongly\nAgree")
                     ) +
  facet_wrap(~factor(vars,
                     levels = c("Binding",
                                "Individualizing"
                                )
                     ), 
             scales = "free_y",
             ncol = 2) + 
  theme(
    plot.title.position = "plot",
    panel.grid.major = element_blank(),
    panel.grid.minor = element_blank(),
    legend.position = "none",
    axis.text.y = element_text(color = "black",
                               size = 11),
    axis.text.x = element_text(color = "black",
                               size = 8,
                               hjust = .5),
    axis.title = element_blank(),
    strip.background.x = element_rect(color = "gray80",
                                      size = 1),
    strip.text.x = element_text(face = "bold")
  )  + 
  labs(
    title = "Endorsement of Individualizing and Binding Moral Values Across Countries",
    caption = "\n\nDashed lines represent the average of all countries.\nColors represent the countries."
    ) 
```

Column {.tabset .tabset-fade data-width=400}
-----------------------------------------------------------------------
### About
This plot visualizes the average scores for individualizing and binding foundations across countries.  

The values on the x-axis range from 0 (i.e., Strongly Disagree) to 5 (i.e., Strongly Agree).  

The final version was improved by (1) modifying the x-axis and (2) explaining the meaning of the colors (i.e., each color represents a country).  

**Descriptive Statistics**  
  
The table below represents the average scores (i.e., mean) for each country as well as standard deviations, minimum and maximum values.  


```{r}
DT::datatable(c_desc,
              colnames = c("Country", "Moral Value",
                           "Mean", "SD", "Min", "Max"),
              rownames = FALSE,
              filter = "none",
              options = list(scrollX = TRUE,
                             pageLength = 6,
                             searching = FALSE,
                             columnDefs = list(list(className = "dt-center",
                                                    targets = "_all")
                                               ),
                             lengthChange = FALSE,
                               initComplete = JS(
                               "function(settings, json) {",
                               "$(this.api().table().header()).css({'background-color': '#9f939c', 'color': '#fff'});",
                               "}")
                             )
              ) %>% 
  DT::formatRound(columns = c("mean", "sd", "min", "max"),
                  digits = 2)
```


### Version 1

```{r}
c_desc %>% 
  ggplot() + 
  geom_col(
    aes(mean, country, fill = vars),
    position = "dodge"
  ) +
  theme(
    panel.grid.major.y = element_blank(),
    panel.grid.minor.y = element_blank()
  ) + 
  scale_x_continuous(expand = c(0, 0))
```

> This is the first version of the plot. There is a lot to work on here. Both axes and legend labels seem confusing. X-axis scale is also not complete. Also, it is hard to see any patterns without using facet_wrap and sorting the values.

### Version 2

```{r}
c_desc %>% 
  ggplot() + 
  geom_col(
    aes(mean, reorder_within(country, mean, vars)
        )
          ) + 
  scale_y_reordered() + 
  scale_x_continuous(expand = c(0, 0)
                     ) +
  facet_wrap(~vars, 
             scales = "free_y",
             ncol = 2) +
  theme(
    panel.grid.major.y = element_blank(),
    panel.grid.minor.y = element_blank(),
    axis.text.y = element_text(color = "black",
                               size = 11),
    axis.text.x = element_text(color = "black",
                               size = 9),
    axis.title = element_blank()
  )  + 
  labs(
    title = "Endorsement of Individualizing and Binding Moral Values Across Countries"
    )
```

> The revised version looks much better. You can see interesting patterns such as Spain's scores. The colors are terrible, though. Also, it'd be nice to see the grand mean to have a general reference category.

### Version 3

```{r}
c_desc %>% 
  ggplot() + 
  geom_vline(data = gmeans,
             aes(xintercept = m), 
             linetype = 2,
             alpha = .6) +   
  geom_col(
    aes(mean, reorder_within(country, mean, vars), 
        fill = country,
        alpha = .9
        )
          ) + 
  scale_y_reordered() + 
  scale_x_continuous(expand = c(0, 0)
                     ) +
  facet_wrap(~vars, 
             scales = "free_y",
             ncol = 2) + 
  theme(
    plot.title.position = "plot",
    panel.grid.major.y = element_blank(),
    panel.grid.minor.y = element_blank(),
    legend.position = "none",
    axis.text.y = element_text(color = "black",
                               size = 11),
    axis.text.x = element_text(color = "black",
                               size = 9),
    axis.title = element_blank()
  )  + 
  labs(
    title = "Endorsement of Individualizing and Binding Moral Values Across Countries",
    caption = "Vertical lines represent the average of all countries."
    ) 
```

> Based on the peer reviews I received, I realized that there is nothing to indicate what the values on the x-axis and the colors mean.


# Predictors of moral values  

Background {.sidebar}
-----------------------------------------------------------------------
**Data**  
Data used in this dashboard come from the second study of [Atari et al. (2020)](http://dx.doi.org/10.1098/rspb.2020.1201), which can be downloaded from [Kaggle](https://www.kaggle.com/tunguz/sex-differences-in-moral-judgements-67-countries).  

This study has data on moral values in 19 countries. There is also a second dataset with country-level variables.

**Country-level variables**  
  
Population Sex Ratio  
Individualism  
Masculinity  
Gender Equality  
Human Development Index  
Overall Life Satisfaction Index  

**Definitions**

Moral foundations theory [(Graham et al., 2013)](https://doi.org/10.1016/B978-0-12-407236-7.00002-4) argues that there are five universal moral values. These five values have been grouped into two overarching moral values [(Graham et al., 2009)](http://dx.doi.org/10.1037/a0015141) as individualizing and binding moral values.  

*Individualizing* values encompass moral foundations of care (i.e., cherishing and protecting others) and fairness (i.e., rendering justice according to shared rules).  

*Binding* values encompass moral foundations of loyalty (i.e., standing with your group, family, nation), authority (i.e., submitting to tradition and legitimate authority), and purity (i.e., abhorrence for disgusting things, foods, actions).  

**Example Items**  

Please read the following sentences and indicate your agreement or disagreement:  

Individualizing: Compassion for those who are suffering is the most crucial virtue.  

Binding: People should be loyal to their family members, even when they have done something wrong.  

Column {data-width=600}
-----------------------------------------------------------------------

```{r MLM, include=FALSE}
# merge country-level data with the main dataset
df <- left_join(df, 
                select(df_c, country:overall_life_satisfaction_index), 
                by = "country")

# MLM
model_i <- lmer(indiv ~ relevel(sex, ref = 2) + # level 1 predictor
                  
                  # level 2 predictors:
                  pop_sex_ratio + individualism + masculinity + 
                  gender_eqality + human_development_index + 
                  overall_life_satisfaction_index + 
                  
                  (sex | country), # random slope for sex
                                   # random intercept for country
                
                data = df
                )

model_b <- lmer(bind ~ relevel(sex, ref = 2) + # level 1 predictor
                  
                  # level 2 predictors:
                  pop_sex_ratio + individualism + masculinity + 
                  gender_eqality + human_development_index + 
                  overall_life_satisfaction_index + 
                  
                  (sex | country), # random slope for sex
                                   # random intercept for country
                
                data = df
                )

# extract coefficients
m_i_fixed <-
broom.mixed::tidy(model_i) %>% 
  filter(effect == "fixed",
         term != "(Intercept)") %>% 
  select(-c(effect, group)) %>% 
  mutate(
         term = recode(term,
                       `relevel(sex, ref = 2)Female` = "Sex [Female]",
                       `pop_sex_ratio` = "Population Sex Ratio",
                       `individualism` = "Individualism",
                       `masculinity` = "Masculinity",
                       `gender_eqality` = "Gender Equality",
                       `human_development_index` = "Human Development Index",
                       `overall_life_satisfaction_index` = "Overall Life Satisfaction Index"),
         model = "Individualizing",
  ) %>% 
  relocate(model, term)

m_b_fixed <-
broom.mixed::tidy(model_b) %>% 
  filter(effect == "fixed",
         term != "(Intercept)") %>% 
  select(-c(effect, group)) %>% 
  mutate(
         term = recode(term,
                       `relevel(sex, ref = 2)Female` = "Sex [Female]",
                       `pop_sex_ratio` = "Population Sex Ratio",
                       `individualism` = "Individualism",
                       `masculinity` = "Masculinity",
                       `gender_eqality` = "Gender Equality",
                       `human_development_index` = "Human Development Index",
                       `overall_life_satisfaction_index` = "Overall Life Satisfaction Index"),
         model = "Binding",
  ) %>% 
  relocate(model, term)

both_ms <- bind_rows(m_i_fixed, m_b_fixed)
```


### **Final Version**

```{r}
# Tried to reorder the y axis, but this didn't work for some reason.
# both_ms %<>% 
#   mutate(
#     term = factor(term, 
#                   levels = c("Sex [Female]",
#                              "Population Sex Ratio",
#                              "Individualism",
#                              "Masculinity",
#                              "Gender Equality",
#                              "Human Development Index",
#                              "Overall Life Satisfaction Index"
#                             )
#                   )
#         )


ggplot() + 
  
  geom_vline(xintercept = 0, 
             colour = "grey60", 
             linetype = 2) +  
  
  geom_errorbarh(data = filter(both_ms, model == "Individualizing"),
                 aes(xmin = estimate + qnorm(0.025)*std.error,
                     xmax = estimate + qnorm(0.975)*std.error,
                     y = term,
                     color = "Individualizing"),
                 position = position_nudge(y = -.1),
                 height = 0
                 ) +
  
  geom_point(data = filter(both_ms, model == "Individualizing"),
             aes(estimate, term, color = "Individualizing"),
             position = position_nudge(y = -.1),
             size = 1.75
             ) +
  
  geom_errorbarh(data = filter(both_ms, model == "Binding"),
                 aes(xmin = estimate - 1.96*std.error,
                     xmax = estimate + 1.96*std.error,
                     y = term,
                     color = "Binding"),
                 position = position_nudge(y = .1),
                 height = 0
                 ) +
  
  geom_point(data = filter(both_ms, model == "Binding"),
             aes(estimate, term, color = "Binding"),
             position = position_nudge(y = .1),
             size = 1.75
             ) + 
  
  scale_color_manual(name = "",
                     values = c("Individualizing" = "cornflowerblue",
                                "Binding" = "#F8766D")
                     ) +
  
  scale_x_continuous(n.breaks = 10) +
  
  theme(plot.title = element_text(vjust = 3),
        plot.title.position = "plot",
        axis.text.y = element_text(color = "black",
                                   size = 11),
        axis.title.y = element_blank(),
        legend.justification = c(0, 0), 
        legend.position = c(.65, .85),
        legend.background = element_rect(colour = "grey80"),
        legend.title = element_blank(),
        panel.grid.major.y = element_blank()
        )  +
  
  labs(title = "Predicting moral values by sex and country-level predictors",
       x = "Unstandardized Coefficient")
  
```

Column {.tabset .tabset-fade data-width=400}
-----------------------------------------------------------------------
### About  

The plot describes the fixed effects for the models described above. The coefficients are unstandardized and represented by the dots. The lines represent the 95% confidence intervals.  

**Findings**  
  
```{r}
table <- 
  bind_cols(
    
    subset(both_ms, model == "Binding",
           select = c(term:std.error, p.value)
           ),
    
    subset(both_ms, model == "Individualizing",
           select = c(term:std.error, p.value)
           )
  ) %>% 
  rename(
    term1 = 1, term2 = 5,
    est1 = 2, est2 = 6,
    se1 = 3, se2 = 7,
    p1 = 4, p2 = 8
  )

table %>% 
  select(-term2) %>% 
  mutate(gr = c("Level 1", rep("Level 2", 6))) %>% 
  group_by(gr) %>% 
  gt() %>% 
  tab_header(
    title = "Predicting Individualizing and Binding Moral Values",
    subtitle = "Across 19 Countries"
  ) %>% 
  cols_label(term1 = "",
             est1 = md("*B*"),
             se1 = "SE",
             p1 = "p",
             est2 = md("*B*"),
             se2 = "SE",
             p2 = "p") %>% 
  tab_spanner(
    label = "Binding",
    columns = vars(est1, se1, p1)
  ) %>% 
  tab_spanner(
    label = "Individualizing",
    columns = vars(est2, se2, p2)
  ) %>% 
  fmt_number(columns = vars(est1, se1,
                            est2, se2),
             decimals = 2) %>% 
  fmt_number(columns = vars(p1, p2),
             decimals = 4
             ) %>% 
  cols_align(
    columns = "term1",
    align = "right"
  ) %>% 
  opt_row_striping() %>% 
  tab_options(
    heading.background.color = "#9f939c",
    table.font.size = "100%"
    ) %>% 
  tab_style(
    style = list(
      cell_text(weight = "bold")
    ),
    locations = cells_body(
      columns = vars(p1),
      rows = p1 < .05
    )
  )
```



### Version 1

```{r}
dwplot(both_ms) + 
  ggtitle("Predicting moral values") +
  xlab("Unstandardized Coefficient") + 
  geom_vline(xintercept = 0, 
             colour = "grey60", 
             linetype = 2)
```

> The initial dot-whisker plot for the fixed effects. Added a vertical line, which made the plot a bit easier to interpret. However, legend looks awful. Modifying the x axis should also help.

### Version 2

```{r}
dwplot(both_ms, 
        dot_args = list(size = 2)
        ) + 
  ggtitle("Predicting moral values by sex and country-level predictors") +
  xlab("Unstandardized Coefficient") + 
  geom_vline(xintercept = 0, 
             colour = "grey60", 
             linetype = 2) + 
  theme(plot.title = element_text(face = "bold"),
        plot.title.position = "plot",
        axis.text.y = element_text(color = "black",
                                   size = 11),
        legend.justification = c(0, 0), 
        legend.position = c(.74, .85),
        legend.background = element_rect(colour = "grey80"),
        legend.title.align = .5
        ) +
  scale_x_continuous(n.breaks = 10) +
  scale_color_manual(values = c("blue", "red"))
```

> Easier to see the points. The legend makes sense now, but still could be better. Colors can be improved. Also, there are too many grid lines.

### Version 3

```{r}
dwplot(both_ms, 
        dot_args = list(size = 2)
        ) + 
  ggtitle("Predicting moral values by sex and country-level predictors") +
  xlab("Unstandardized Coefficient") + 
  geom_vline(xintercept = 0, 
             colour = "grey60", 
             linetype = 2) + 
  theme(plot.title = element_text(face = "bold", vjust = 3),
        plot.title.position = "plot",
        axis.text.y = element_text(color = "black",
                                   size = 11),
        legend.justification = c(0, 0), 
        legend.position = c(.65, .85),
        legend.background = element_rect(colour = "grey80"),
        legend.title = element_blank(),
        panel.grid.major.y = element_blank()
        ) +
  scale_x_continuous(n.breaks = 10) +
  scale_color_manual(values = c("cornflowerblue", "#F8766D"))
```

> Based on the peer review I received, I realized that the order of the variables in the legend is reversed.

# Sex X Moral Values for each country

Background {.sidebar}
-----------------------------------------------------------------------
**Data**  
Data used in this dashboard come from the second study of [Atari et al. (2020)](http://dx.doi.org/10.1098/rspb.2020.1201), which can be downloaded from [Kaggle](https://www.kaggle.com/tunguz/sex-differences-in-moral-judgements-67-countries).  

This study has data on moral values in 19 countries. There is also a second dataset with country-level variables.

**Country-level variables**  
  
Population Sex Ratio  
Individualism  
Masculinity  
Gender Equality  
Human Development Index  
Overall Life Satisfaction Index  

**Definitions**

Moral foundations theory [(Graham et al., 2013)](https://doi.org/10.1016/B978-0-12-407236-7.00002-4) argues that there are five universal moral values. These five values have been grouped into two overarching moral values [(Graham et al., 2009)](http://dx.doi.org/10.1037/a0015141) as individualizing and binding moral values.  

*Individualizing* values encompass moral foundations of care (i.e., cherishing and protecting others) and fairness (i.e., rendering justice according to shared rules).  

*Binding* values encompass moral foundations of loyalty (i.e., standing with your group, family, nation), authority (i.e., submitting to tradition and legitimate authority), and purity (i.e., abhorrence for disgusting things, foods, actions).  

**Example Items**  

Please read the following sentences and indicate your agreement or disagreement:  

Individualizing: Compassion for those who are suffering is the most crucial virtue.  

Binding: People should be loyal to their family members, even when they have done something wrong.   


Column {data-width=600}
-----------------------------------------------------------------------

### Final Version

```{r}
ms1 <- lmer(indiv ~ sex + (sex|country), data = df)
ms2 <- lmer(bind ~ sex + (sex|country), data = df)

p1 <- ggpredict(ms1,
                terms = c("sex", "country"),
                type = "re"
                )

p2 <- ggpredict(ms2,
                terms = c("sex", "country"),
                type = "re"
                )

p1 %<>% 
  as.data.frame() %>% 
  mutate(
    outcome = "Individualizing"
  )

p2 %<>% 
  as.data.frame() %>% 
  mutate(
    outcome = "Binding"
  )

p_combined <- 
  bind_rows(p1, p2) %>% 
  rename(Sex = x)

ggplot(p_combined) + 
  geom_line(aes(predicted, group),
            size = 2.5,
            color = "#535353") + 
  geom_point(aes(predicted, group, color = Sex),
             size = 4) + 
  geom_hline(yintercept = 0.5,
             color = "gray90") +
  facet_wrap(~outcome) +
  theme(
    plot.title.position = "plot",
    panel.grid.major.y = element_blank(),
    panel.grid.minor.y = element_blank(),
    legend.position = c(0.1, -.2),
    legend.direction = "horizontal",
    axis.text.y = element_text(color = "black",
                               size = 11),
    axis.text.x = element_text(color = "black",
                               size = 8,
                               hjust = .5),
    axis.title.y = element_blank(),
    axis.title.x = element_text(vjust = -2),
    strip.background.x = element_rect(color = "gray80",
                                      size = 1),
    strip.text.x = element_text(face = "bold")
  )  + 
  labs(
    title = "Predicted values of Individualizing and Binding Moral Values by Sex",
    caption = "\nDashed lines represent the average predicted values of all countries.",
    x = "Predicted Values"
    ) +
  scale_color_manual(values = c("#95c3db", "#1a64a4"))
```


```{r}
# run the models
model1 <- lmer(indiv ~ gender_eqality + (sex|country), data = df)
model2 <- lmer(bind ~ gender_eqality + (sex|country), data = df)

# extract the predicted values
predicted1 <- 
  ggpredict(model1, 
            terms = c("gender_eqality", "country"),
            type = "re")

predicted2 <- 
  ggpredict(model2, 
            terms = c("gender_eqality", "country"),
            type = "re")

predicted1 %<>% 
  as.data.frame() %>% 
  mutate(
    outcome = "Individualizing"
  )

predicted2 %<>% 
  as.data.frame() %>% 
  mutate(
    outcome = "Binding"
  )

pred_combined <- bind_rows(predicted1, predicted2)

pred_combined %>% 
  ggplot(aes(x, predicted, color = group)) +
  geom_line(size = .75) + 
  facet_wrap(~ outcome) +
  labs(title = "Relationship between gender equality and moral values",
       x = "Gender Equality",
       y = "Predicted Values",
       color = "Country") +
  theme(
    plot.title.position = "plot",
    axis.text = element_text(size = 10,
                             colour = "black"),    
    strip.background.x = element_rect(color = "gray80",
                                  size = 1),
    strip.text.x = element_text(face = "bold",
                                size = 11)
        
        ) +
  gghighlight(group %in% c("Japan", "Netherlands", "Spain", "Iran"),
               calculate_per_facet = TRUE)
```

Column {.tabset .tabset-fade data-width=400}
-----------------------------------------------------------------------

### About

The plot describes the fixed effects for the models described above. The coefficients are unstandardized and represented by the dots. The lines represent the 95% confidence intervals.  

**Findings**  
  
```{r}
table <- 
  bind_cols(
    
    subset(both_ms, model == "Binding",
           select = c(term:std.error, p.value)
           ),
    
    subset(both_ms, model == "Individualizing",
           select = c(term:std.error, p.value)
           )
  ) %>% 
  rename(
    term1 = 1, term2 = 5,
    est1 = 2, est2 = 6,
    se1 = 3, se2 = 7,
    p1 = 4, p2 = 8
  )

table %>% 
  select(-term2) %>% 
  mutate(gr = c("Level 1", rep("Level 2", 6))) %>% 
  group_by(gr) %>% 
  gt() %>% 
  tab_header(
    title = "Predicting Individualizing and Binding Moral Values",
    subtitle = "Across 19 Countries"
  ) %>% 
  cols_label(term1 = "",
             est1 = md("*B*"),
             se1 = "SE",
             p1 = "p",
             est2 = md("*B*"),
             se2 = "SE",
             p2 = "p") %>% 
  tab_spanner(
    label = "Binding",
    columns = vars(est1, se1, p1)
  ) %>% 
  tab_spanner(
    label = "Individualizing",
    columns = vars(est2, se2, p2)
  ) %>% 
  fmt_number(columns = vars(est1, se1,
                            est2, se2),
             decimals = 2) %>% 
  fmt_number(columns = vars(p1, p2),
             decimals = 4
             ) %>% 
  cols_align(
    columns = "term1",
    align = "right"
  ) %>% 
  opt_row_striping() %>% 
  tab_options(
    heading.background.color = "#9f939c",
    table.font.size = "100%"
    ) %>% 
  tab_style(
    style = list(
      cell_text(weight = "bold")
    ),
    locations = cells_body(
      columns = vars(p1),
      rows = p1 < .05
    )
  )

pred_combined %>% 
  arrange(group) %>% 
  gt()
```

### Version 1

```{r}
# plot
p1 <- 
predicted1 %>% 
  ggplot(aes(x, predicted, color = group)) +
  geom_line(size = 1)

p2 <- 
predicted2 %>% 
  ggplot(aes(x, predicted, color = group)) +
  geom_line(size = 1)

ggpubr::ggarrange(p1, p2,
                  common.legend = TRUE,
                  legend = "bottom"
                 ) 
```

> The initial attempt to visualize the predicted values of moral foundations where the predictor is gender equality. I chose this predictor because it was the only that was significant for both outcomes. I am glad that it worked, but it needs improvement.

### Version 2

```{r}
p1 <-
p1 + 
  labs(x = "Gender Equality",
       y = "Individualizing",
       color = "Country") +
  theme(axis.text = element_text(size = 10,
                                 colour = "black"))

p2 <-
p2 + 
  labs(x = "Gender Equality",
       y = "Binding",
       color = "Country") +
  theme(axis.text = element_text(size = 10,
                                 colour = "black"))

ggpubr::ggarrange(p1, p2,
                  common.legend = TRUE,
                  legend = "bottom"
                 ) 
```

> This version is better, but it would be nicer to see which countries are at the top, middle, and bottom of the plot.


### Version 3

```{r}
# run the models
model1 <- lmer(indiv ~ gender_eqality + (sex|country), data = df)
model2 <- lmer(bind ~ gender_eqality + (sex|country), data = df)

# extract the predicted values
predicted1 <- 
  ggpredict(model1, 
            terms = c("gender_eqality", "country"),
            type = "re")

predicted2 <- 
  ggpredict(model2, 
            terms = c("gender_eqality", "country"),
            type = "re")

p11 <- 
  predicted1 %>% 
  ggplot(aes(x, predicted, color = group)) +
  geom_line(size = 1) +
  labs(x = "Gender Equality",
       y = "Individualizing",
       color = "Country") +
  theme(axis.text = element_text(size = 10,
                                 colour = "black")
        ) +
  gghighlight::gghighlight(group %in% c("Poland", "Netherlands", "Hungary"))

p21 <- 
  predicted2 %>% 
  ggplot(aes(x, predicted, color = group)) +
  geom_line(size = 1) +
  labs(x = "Gender Equality",
       y = "Binding",
       color = "Country") +
  theme(axis.text = element_text(size = 10,
                                 colour = "black")
        ) +
  gghighlight::gghighlight(group %in% c("Poland", "Netherlands", "Spain"))

ggpubr::ggarrange(p11, p21,
                  common.legend = TRUE,
                  legend = "bottom")  %>%  
  ggpubr::annotate_figure(
    top = ggpubr::text_grob("Countries at the top, middle, and bottom")
    )
```